EU’s AI Act - not only about AI, but is it also about ... ESG?
Although the title of the EU’s AI Act regulation suggests that it will strictly regulate issues related to the creation and use of artificial intelligence, with areas such as environmental protection remaining outside its scope - nothing could be further from the truth.
On August 1, 2024, the AI Act, the world's first such comprehensive legal regulation of artificial intelligence, went into effect. Furthermore, because of the AI Act’s comprehensiveness, environmental aspects cannot be omitted.
Why is the AI Act also about environmental protection?
The fundamental purpose of the regulation is to ensure the development of artificial intelligence that is human-centered (human-centric approach) and trustworthy - with respect to fundamental rights. And one of these rights (along with health and safety) is the protection of the environment. The Charter of Fundamental Rights makes it clear that "a high level of environmental protection and the improvement of its quality must be integrated into the Union's policies and ensured in accordance with the principle of sustainable development" (Article 27).
Hence, among the obligations imposed on developers, suppliers and users of AI systems also include requirements relating to environmental protection.
What are these requirements?
The entire regulation is based on adjusting specific measures in proportion to the risks that, according to an assessment, the use of a given system may pose (including to the environment) Although of course, most rigor applies to high-risk systems, guidelines aimed at the Ai issue will certainly serve as best practice for other areas as well. The requirements apply not only to the design, development, documentation and implementation of systems - but also to their monitoring and incident reporting as responsibilities arise at every stage of the system life cycle. Importantly, a major incident is also considered to be the one that leads to, may have led to, or may directly lead to environmental harm.
On the other hand, the positive impact of creating solutions based on artificial intelligence that enable better forecasting, optimization of operations, and allocation of resources - for the benefit of resource and energy efficiency and climate change mitigation and adaptation - is also appreciated.
And from yet another perspective - let's not forget how much computing power (and thus energy) is required for the creation and utilization of such systems.
Given the above, how can one meet the requirements of #ESG?
An attempt to answer this challenge is green software development, an approach that focuses on creating applications and IT systems in a sustainable and environmentally friendly way. This can be done by, among other things
- optimizing code to reduce energy consumption during application operations,
- choosing energy-efficient coding languages, such as Python, instead of more energy-intensive languages, such as C,
- minimizing the use of hardware resources,
- designing "lightweight applications" with a smaller digital footprint,
- using cloud computing to optimize server utilization and integrating RES in data centers,
- creating user interfaces that promote energy savings,
- implementation of energy-saving functions in applications,
- but also continuous monitoring and optimization of environmental performance, using tools to measure energy consumption and CO2 emissions associated with software.
Therefore, various tools can be used to help create more sustainable and energy-efficient software, but environmental aspects at all stages of development must be taken into account, from design to implementation and maintenance - throughout the lifecycle of the product. The process is not about necessarily achieving a zero-carbon standard - but about a series of small actions implemented by teams that, in effect, minimize the digital footprint of the product.
It's also worth noting that, given the difficulty of measuring and precisely defining savings, there is a risk of greenwashing - but this is a problem we will generally face with all ESG compliance.
Author: Magdalena Raniszewska, Legal Counsel and Managing Partner at RBR Law Firm, which is ABSL's Regional Legal Partner in the Tricity Chapter